Annual Meeting of the Rapaport-Klein Study Group
AustenRiggsCenter, Stockbridge, Massachusetts,June 3-5, 2011
Response to commentaries by David Wolitzky, Morris Eagle and Jeremy Safran
at the 2010 Rapaport-Klein Study Group Meeting
Irwin Hoffman
Third of four files:
Interpolated comments to the paper by Morris N. Eagle and David L. Wolitzky,
“Systematic Empirical Research versus Clinical Case Studies: A Valid Dichotomy?”,
Journal of the American Psychoanalytic Association, 2011, Vol. 59, in press
(Part 2)
Part II – The Case Study Method
From its inception (throughout most of the 20th century,) the psychoanalytic case study has enjoyed privileged status, vis a vis, systematic empirical research, as the mean s of establishing and advancing psychoanalytic knowledge.
We know that most case reports consist of vignettes selected to support a hypothesis I’ve never written about a case to “support a hypothesis.” I write to convey how various factors are woven together to form a complex whole that contributes to therapeutic action. The list of factors involved in the process is long. I’m not interested in how any one of them “worked” but in an overall multifaceted story. Isolating one or two variables would do violence to the sense of the whole fabric of the analytic relationship that I want to convey. I do not think you could read any of my reports of clinical experience and comfortably formulate a hypothesis that it seems to “generate” and that one could then proceed to “test.” I think most case reports that I read are like that.rather than beinga complete and faithful account of what transpired. “A complete and faithful account” is not possible, especially if you mean that in the sense of the objective “facts.” The data are ambiguous. Any “account” must be through a lens, organized according to a perspective or set of perspectives. A different lens might bring out other things. The number of lenses through which the data could be viewed is infinite just as the number of plausible percepts of an area of a Rorschach card is infinite. A tape recording does not yield all the relevant data because the participants’ running experiences are integral to the process and they are not visible. Those experiences, conscious and unconscious, and the whole history of the relationship create the context for anything that happens in a current interaction, for every word that is spoken. Methods applied to the data to capture what “really” went on will only yield a picture that is relative to those methods.Thus, years after Freud expressed his concerns, Anna Freud (1971. p. ix) implied a similar uneasiness when she noted that “We cannot help being conscious ... of a conspicuous ... dearth of ... complete Impossible. No such thing exists. As Merleau-Ponty says (I quote him in a couple of papers on this) it’s not because of the limitations of the knower, it’s because of the nature of the object, the “structure of human experience.”and adequately documented case histories.” As Michels (2000) pointed out, a survey of the psychoanalytic literature from 1969–1982 that focused on the articles cited most frequently apparently failed to find any extensive case study reports (Klumpner & Frank, 1991). Other analysts, however, seem to feel that relying on selected case vignettes is fine, indeed preferable to full-length reports because they provide a more vivid account of the analytic work (e.g., Stein, 1988). I’d agree. The aim is to heighten people’s consciousness of possibilities to consider, and vignettes selected for the purpose of illustration work fine for that purpose.
In relation to the issue of selectively regarding which aspects of which cases are found in the literature, Michels (2000) invites us to pay attention to the analyst’s purposes in writing up a case and publishing it. He indicates that when the intention is to offer evidence for an analytic hypothesis about the meanings of some aspect of the patient’s behavior, I don’t have much time to read altogether, but the case reports I’m familiar with just don’t have that kind of structure. At least it’s not what I see. Maybe it’s because I am not very likely to read cases that are by an avowed advocate of a specific theory. I don’t often read cases by self psychologists or Kleinian analysts for example. I might read some relational case studies but I think of “relational-constructivist” as a meta-perspective, not a theory about specific motives. many observers believe it would be useful to have a tape and a transcript. On, the other hand, as Michels (2000) notes, Galatzer-Levy (1991. p. 736), in a panel report of the Committee of Scientific Activity (of the American Psychoanalytic Association), comments that the preference for verbatim data is “scientism,” “... the irrational veneration of what appears scientific rather than using scientific methods as tools.” He states “Abandoning narratives would deprive us of the richly informative narrator's perspective.” This view represents an unnecessary choice. It need not be ‘either-or’. Obviously, the narrator’s perspective can be “richly informative,” and would be even more informative if it was accompanied by a record of the thoughts and feelings the analyst experienced during the sessions on which the narrative is based. If the intention is to get something “comprehensive,” this seems, on the contrary, like a tedious chore yielding extremely selective, possibly contrived accounts. I’d rather have the analyst speak of what he or she actually felt was important. In effect I’d rather the selectivity be in the open along with whatever the analyst can say about his or her biases. In effect, I think the whole strategy that you advocate, searching for objective-comprehensive-original data encourages bias that is disclaimed and subjectivity that is disclaimed,At the same time, the “richness” would be enhanced by also having the verbatim material for others to study in a systematic fashion. In fact, comparing the analyst’s narrative with what might emerge from a detailed study of the original data by independent observers could be quite illuminating and more “richly informative” than either source of data alone. One thing that strikes me is that this kind of report would be extremely boring, I mean really deadly. Life is short. There’s no way I could bear reading anything like that. Intrinsic to the boredom I think is the reader’s sense that what will be learned from all of this kind of gathering of “unbiased” data will be extremely thin at best. Such an approach would reduce the common limitations of case studies: 1) distortion of case material and/or facts in the patient’s history in the service of presenting a more compelling set of assertions; 2) unwitting distortion or selective memory of facts and/or clinical data in the service of offering a more persuasive case or as a result of countertransference reactions; 3) deliberate disguise of the patient’s identity that results in the alteration of clinical data or facts about the patient’s history that others might feel renders some of the inferences drawn questionable. There should be a comparable list of “distortions” that could emerge from the study of “verbatim material” in a “systematic fashion.” There’s no way to “study” without applying “methods” and those methods will be selective in what they capture, even, according to some other points of view, selective to the point of leaving out the most significant things.
The problem of “confirmatory bias”
A systematic, empirical approach might shed light on the issue of biased weighting of clinical evidence. In this regard, one of us participated in a research project on clinical evidence in which several analysts studied the verbatim transcripts of numerous analytic sessions. The group, organized and led by Benjamin Rubinstein., met regularly. We started by reading the transcripts of the first five sessions. Any time a member of the group had a hypothesis to offer, we stopped and recorded the hypothesis and the observations on which it was based. In subsequent meetings, we read transcripts of randomly selected subsequent sessions. When a group member felt there was evidence, either in favor of, or against, a given hypothesis, we stopped and rated the strength of the evidence. Two noteworthy findings emerged from this procedure. First, 98% of the ratings were in the positive direction, meaning that we rarely regarded a hypothesis to have been disconfirmed by the clinical material. Second, when we compared the ‘strength of evidence’ ratings of the person who had us stop to rate the evidence for a given hypothesis and compared that rating to the average the rating of the other group members (i.e., those who did not call attention to clinical material), the group rating was lower, with the exception of one analyst (out of the eight in the research group).
What this analysis suggests is that the analyst who felt there was evidence for a hypothesis (which did not necessarily have to be the one he himself proposed originally) thought the evidence was stronger than did his colleagues. In short, there was an indication of what we might call a “confirmatory bias,” defined above, and expressed in our group by the tendency to give more weight to evidence than other colleagues feel is warranted, Another noteworthy finding is that it was quite rare (i.e., less than 5% of the time) for anyone to find negative evidence of any initial hypothesis. This finding is somewhat ambiguous in that it could reflect a confirmatory bias or the extraordinary clinical acumen of the clinicians! Extrapolating from these findings to the clinical situation, it is likely that 1) we rarely regard our initial hypotheses as disconfirmed or as not supported by further clinical observations, and, 2) we give greater weight to apparently “confirmatory” evidence than is warranted. How does one decide what is “warranted”? It seems reasonable to regard this as a limitation of the case study method. At the very least, this kind of ‘confirmatory bias’ suggests room for improvement in the processing and reporting of case material. There is no “objectively” correct weight that some piece of the analytic interaction should have. A theoretical or personal bias might highlight certain things that would not be as salient with a different bias. A reporting analyst conveys enough that is amenable to many plausible interpretations so that constructive discussion of pros and cons of multiple points of view can ensue. You seem to be looking for ways to decide definitively which point of view is best. I think that quest is futile and discouraging of ongoing openness and reflectiveness. One person could see something that is brilliant and original and that some others might find convincing even though they might not have seen it until it was brought to their attention. Nevertheless it could be superseded by another view that emerges later. The analyst might be one of those with an especially illuminating perspective. In a broad sense we want to encourage the patient to consider various possibilities and to cultivate a mindset that is reflective and open in an ongoing way The ideal climate of the “conversation” of the analyst and the patient is similar or the same as the ideal climate of discussion and debate in the analytic community. There is literature (Habermas, Gadamer) on the nature of such dialogue that analysts (myself included) would do well to study. It is all what I would call “constructivist” or “critical constructivist” or, to use my own term, “dialectical constructivist.”
Suggestions for improving the quality and bolstering the evidential value of case studies
Even if one wants to maintain that case studies are all that we need in psychoanalysis, Hoffman’s paper is silent on the question of the adequacy of the typical case study (or clinical vignette) found in the analytic literature. I do not see how you have taken on that challenge any more than I have. What grounds do you have for making claims about what is or is not “typical.” There are thousands of case studies and I don’t think it makes sense to call something “typical.” I would think the thing to do is give an example of a case study that has features you find objectionable and another with features that you think are predominantly good.If he wants to persuade us that such studies are to be preferred over research studies as the basis for psychoanalytic knowledge, I think this is a sloppy representation of my argument. At the core of it is opposition to the privileging of systematic empirical research, which you announce you favor at the outset (with no caveats, by the way, about qualities or purposes of systematic research or case studies) when you speak of the motivation for the “standing ovation.” I believe I bring to bear very rigorously the specific strengths of the case study method and the weaknesses of systematic research that bear on this issue. And it won’t surprise you that I think those strengths and weaknesses add up to an overwhelmingly compelling argument that the privileging, the hierarchical arrangement of hypothesis-testing research vis-à-vis “anecdotal” hypothesis-generating case studies, is unjustified. It’s in the specific nature of the argument that I feel make it not susceptible to being overridden by consideration of the weaknesses of the case study method or the strengths of systematic research (Of course we also clearly do not agree about what those are).one would have hoped that he might spell out the criteria for case studies that should be regarded as yielding reliable knowledge. It yields possibilities to consider. How “reliable” does it have to be to be trusted for that?
Case studies should be accorded more evidential value to the extent that they demonstrate the following characteristics:
1) The ratio of theory to data is reasonable, i.e., there is not an excessive amount or theory superimposed on some fragment of data.
2) Observation is clearly separated from inference in the case report. I think that’s not possible and that it reflects a positivist epistemology, the claim that there is such a thing as pure observation without inference. Can you describe any affective state without inference? Consider for example my account of my experience with Ken at the elevator. At the end of a session, the first time in my private office on a high floor [he’s phobic of heights], he asks me to walk with him to the elevator.I responded to Ken's request immediately simply by saying "sure" and we walked to the elevators. My immediate feeling was that it would have been extremely stingy of me to decline or even to hesitate since it had been such an ordeal for Ken to tolerate the session in this office. I knew, after all, that the idea of meeting at this location was initiated originally by me. Also, the patient's request, an aggressive initiative on his part, was out of character. It was a risk for him to make it and I thought he might well feel, not only disappointed but also humiliated if I said no. I certainly didn't want to be like his father blocking his shots in basketball. That danger seemed greater to me than the dangers of complying. Also, because the request was so unusual, I felt inclined to give the patient the benefit of the doubt and respect whatever creative wisdom might have prompted it. Another consideration might have been that I felt that over time I had conveyed enough of an impression of personal availability to contribute to the patient's readiness to make the request. In any case, as Ken and I waited in the hallway we made a little small talk about the elevators, the express type versus the local type, which stopped at which floors, which he came up on, and so on. After a couple of minutes one opened up and Ken stepped in. We shook hands just as the doors began to shut. It was not our customary way of parting. I'm not sure which of us reached out first. Before getting to the patient's retrospective view of the experience the next day I want to stop to talk a bit more about the episode at the elevator, an example of an" extra-analytic" interaction. How do we conceptualize the nature of the interaction in the hallway? On the surface it could hardly be more mundane. Just a little rather uninteresting small talk. But as we are waiting there is a little tension in the air, a touch of awkwardness, and a feeling that what's happening has a little extra "charge." Would we say that the analyst, ideally, would feel entirely comfortable in that situation? Would we say that the patient too would be comfortable the closer he was to completing his analysis? My own view is that regardless of the specific personalities of the participants and regardless of the amount and quality of analytic work each has under his or her belt there is a residue of tension that is likely because here, in the hallway, outside the psychoanalytic routines of time, place, and role-defined interactions, the analyst emerges out of the shadows of his or her analytic role and is exposed more fully than usual as a person like the patient, as a vulnerable social and physical being. (R and S, pp. 232-233). This account is just an alloy of “observation” and “theory” in a way that I think does not allow for separating them. It’s an “alloy” or a “compound” not a “mixture.” There is so much context, reference to the patient’s history, reference to my behavior in the past, reference to my countertransference, reference to the unusual nature of this request which might be associated with its “creative wisdom.” Even that idea is informed by a special interest that I have in recognizing the patient as a creative agent in the process. What are the junctures in this account where you want independent judges to have a video (unrealistic to be sure, but even if it were possible) and to decide that what I say is or is not “distorting” of what “really” happened? In fact, objectively, was there “a little tension in the air, a touch of awkwardness, and a feeling that what’s happening has a little extra ‘charge’”? And make no mistake, I am telling this story, in fact I even notice it happening as a significant thing, because of a preconception that I have about the “dialectic of ritual and spontaneity in the psychoanalytic process.” It’s also saturated with sensitivity to the issue of “repetition” versus “new experience.” That is theory too. Every word is theory and every word is observation. What would inference-free observation look like? Wouldn’t it have to be devoid of any “meaning”? I moved my head a little to the left. I mouthed the words “the elevators are slow”? He said “uh huh”? Let an independent judge decide whether there really was a “touch of awkwardness”? For me, all the things you talk about have virtually nothing to do with the phenomena that I am interested in. You really want to invest thousands of dollars and a huge amount of manpower to have independent judges decide what “ really” happened in my walk with Ken to the elevator and countless other episodes?